I was on a 6-person app-dev team where the lead wanted to split a new feature into separate microservices right away. I worried about added operational cost and rollout risk for low initial traffic, so I did a 3-day spike: built a modular-monolith prototype, ran basic load tests, and produced an infra cost comparison. The numbers were clear—the microservice route would add about $1,200/month (≈35% higher) and 2–3 extra weeks of integration work. I demoed the prototype and shared the cost model, then proposed starting with a modular monolith with well-defined interfaces so services could be extracted later. We adopted that compromise, shipped in 6 weeks, saved roughly $14,400 in year-one costs, and later extracted a service when traffic justified it. I learned to use quick prototypes and business metrics to influence technical decisions.
Get AI-powered feedback on your answer and improve your skills
Takes 5-10 minutes